Establishment of a relationship between landslide susceptibility zonation and threshold rainfall intensities

Abstract

Landslides induced by heavy rainfall pose a common geotechnical hazard in Sri Lanka. Sloping grounds in Sri Lanka are made of rocks at different levels of weathering, residual and colluvial soils. The parent rock's mineralogical composition influences the heterogeneous nature of these weathering profiles. Slopes in the region have deep groundwater tables and unsaturated soil profiles with high matric suction during dry periods. However, prolonged rainfall leads to slope instability due to the loss of matric suction and the potential formation of a perched water table or the rise of the groundwater table. National Building Research Organisation (NBRO) has developed the landslide hazard zonation maps of Sri Lanka at 1:50000 and 1:10000 for districts where landslide risk is high based on the overall hazard rating (score) determined by terrain factors. The terrain factors are bedrock geology and geological structures, type and natural soil cover and thickness, slope range and category, hydrology and drainage, land use and management and landform. In landslide hazard zonation maps, sloping grounds are divided into regions of different hazard levels: Safe slope, landslide not likely to occur, a modest level of landslide hazard and landslides are to be expected. To supplement that zonation, it is essential to identify the threshold rainfall intensities causing a disastrous situation in zones of different landslide susceptibility. Currently, the threshold rainfall values are determined by previous experience. Researchers have developed different empirical statistical relationships to identify threshold intensities that vary with local landslide factors and site-specific. Given the significant material variations in Sri Lankan slopes, relying solely on statistical approaches is inappropriate. Landslide susceptibility level-specific thresholds based on numerical studies would be necessary to make reliable decisions regarding early warning and remedial measures. In this research, landslides where the triggering rainfall data is available are back analysed to establish a relationship between landslide susceptibility zonation and threshold rainfall intensities. The SEEP/W 2018 software was used to model rainwater infiltration and the consequent changes in the pore water pressure. The results of the above seepage analysis were then transferred to SLOPE/W software to analyse the slope stability. Initially, recent landslides (Pinnawala landslide and Kithulgala landslide) where rainfall records are available was back analysed and threshold rainfall values for these two slopes was established. The level of hazard of these two landslide areas was identified using the landslide hazard zonation maps of Sri Lanka that developed by National Building Research Organisation (NBRO). A relationship between landslide susceptibility zonation and threshold rainfall intensities can be established by doing this analysis on several landslides that rainfall data was available.